Robert A. Laird and Brandon S. Schamp. 2008. Does local competition increase the coexistence of species in intransitive networks? Ecology 89:237–247.

Appendix C. Discussion of an alternative explanation for the reversal in the effect of local vs. global intransitive competition on short- and long-term coexistence.

An alternative explanation for the reversal in the effect of local versus global intransitive competition on short- and long-term coexistence is that while extinction initially occurs relatively slowly in locally competing communities, it may lead to progressively depauperate communities with species compositions that are more prone to extinction than those that arise in globally competing communities. This positive feedback mechanism could occur if, for example, the spatial aggregation of locally competing species leads to fewer interspecific boundaries, resulting in greater variation in the order that species go extinct. In this case, the typical extinction sequence should be (1) more idiosyncratic and (2) more likely to lead to extinction-prone hierarchies, in locally competing communities than in globally competing communities. Using our additional s = 19 model runs, for which we recorded competitive rank (i.e., of the row sums in the competition matrix) and extinction order, we found that the first of these "extinction sequence" explanations was not supported. For every model run in which an extinction occurred, competitive rank was negatively correlated with the extinction order; however, there was no difference in the strength of these correlations between local and global competition (t test: t = −0.96, P = 0.34, N_L = N_G = 24). Thus, species were no more likely to go extinct in an idiosyncratic order in locally- vs. globally competing communities. The second "extinction sequence" explanation also garnered little support. Using our original data set for 4 to 25 species, we found that the change in intransitivity after a community’s first extinction event was generally not significantly different between locally- and globally competing communities (Wilcoxon rank sum tests: P > 0.05 for 20 of 22 comparisons). Overall, these results strongly support the explanation that intransitive, globally competing communities experienced greater long-term coexistence due to the escape of competitively inferior species, and not because the extinction sequence for locally competing species led to depauperate communities whose competitive topologies were more prone to future extinction events.